with the skill
Reverting the headline on this evidence would be a guess. Three things changed, and the headline is probably the smallest contributor.
**Decomposing the 30% drop**
- Traffic fell 22%, so with an unchanged conversion rate, signups would have fallen about 22%.
- Actual signups are at 70% of last week, versus the 78% you'd expect from traffic alone.
- That implies conversion rate fell about 10% relative (0.70 / 0.78 ≈ 0.90).
So roughly three-quarters of the drop is a traffic problem, and the paid pause is the obvious cause (Occam's Razor). The headline can only account for part of the remaining ~10%, and the phone field is a competing suspect for that same slice.
**Complications**
- **Traffic mix changed.** Paid traffic usually converts differently from organic or direct. If paid converted worse, pausing it should have raised your conversion rate, which would make the underlying drop larger than 10%. If it converted better, the drop is smaller. You can't interpret the conversion rate until you segment by source.
- **The phone field is a classic conversion killer.** Required phone numbers commonly cause meaningful drop-off, since they raise perceived risk and effort (Hick's Law, activation energy). It hit mid-week, so it affected only part of the week.
- **Fundamental attribution error.** "People hate the new message" blames the customer's reaction, when the situational causes (fewer visitors, more form friction) are more likely.
**What I'd check, in order**
1. **Conversion by day.** Monday is your cleanest headline-only read, since paid was still running and there was no phone field. Tuesday adds the paid pause. Wednesday onward adds the phone field. If Monday's rate looked normal, the headline is mostly cleared. Daily numbers are noisy, so treat this as directional.
2. **Conversion by traffic source,** especially organic and direct only, before versus after.
Show the rest (13 more lines)
3. **Form funnel:** the share of people who start the form and submit it, before versus after Wednesday. If submission rate fell, it's the phone field.
4. **Homepage-to-signup-page click-through.** If the headline is the problem, it should show up here, before the form.
5. **Whether paid traffic landed on the homepage** or a dedicated landing page. That tells you how much of the 22% traffic drop paid explains.
**What I'd do now**
- Make the phone field optional or remove it. It's cheap, reversible, and a likely culprit.
- Get finance's decision on paid so you can restore that traffic.
- Keep the new headline for now, and don't revert everything at once. That would confound the data again. If you still doubt the headline after the segmented data, run a proper A/B test.
**Ground rule for next time:** change one conversion-critical thing at a time, or at least log changes and keep a holdout. Three simultaneous changes made this week unreadable (Map ≠ Territory: your dashboard shows a drop but not its cause).
If you can share daily signups, visits, and form-start numbers, I can help narrow it down further.